jialicheng/imdb-bert-large

bert-large

This model is a fine-tuned version of bert-large-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9621
• Accuracy: 0.8887

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 32
• eval_batch_size: 256
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	782	0.2848	0.8852
0.3133	2.0	1564	0.3038	0.8888
0.1751	3.0	2346	0.5035	0.8791
0.1057	4.0	3128	0.5942	0.885
0.1057	5.0	3910	0.5220	0.8764
0.0733	6.0	4692	0.6981	0.8823
0.0439	7.0	5474	0.6775	0.8833
0.0371	8.0	6256	0.6118	0.8891
0.0277	9.0	7038	0.7128	0.8864
0.0277	10.0	7820	0.7555	0.8868
0.0202	11.0	8602	0.7618	0.8888
0.0141	12.0	9384	0.7654	0.8842
0.0125	13.0	10166	0.8345	0.8867
0.0125	14.0	10948	0.8073	0.8844
0.0077	15.0	11730	0.7047	0.8887
0.0071	16.0	12512	0.8622	0.8891
0.004	17.0	13294	0.8655	0.8900
0.0031	18.0	14076	0.9096	0.8898
0.0031	19.0	14858	0.9454	0.8892
0.0016	20.0	15640	0.9621	0.8887

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.15.2